Transipedia.org: k-mer-based exploration of large RNA sequencing datasets and application to cancer data

Table 3 Accuracy measures of Reindeer mutation and fusion calls

		#total probes	#probes after selection^a		min_hits^c=1	min_hits=3	min_hits=5	min_hits=10
Mutations	All Depmap mutations^b (50 cancer genes)	3685	3621	^dTrue +	4346	4255	4205	4026
				False +	11810	911	589	484
				False -	255	346	396	575
				Precision	0.269	0.824	0.877	0.893
				Recall	0.945	0.925	0.914	0.875
	Cosmic Hotspot mutations	960	951	True +	1665	1631	1611	1558
				False +	6823	184	114	90
				False -	51	85	105	158
				Precision	0.196	0.899	0.934	0.945
				Recall	0.970	0.950	0.939	0.908
Fusions	All Depmap fusions (junction at exon edges)	8972	8860	True +	9410	9277	9201	9018
				False +	25732	10048	6558	2378
				False -	170	303	379	562
				Precision	0.268	0.480	0.584	0.791
				Recall	0.982	0.968	0.960	0.941
	Cosmic fusions	60	59	True +	99	98	98	96
				False +	22	3	3	2
				False -	1	2	2	4
				Precision	0.818	0.970	0.970	0.980
				Recall	0.990	0.980	0.980	0.960

^aMin count cutoff and low complexity masking
^bRestricted to RNA-seq-derived mutations
^cmin_hits: minimum number of positive k-mers in query
^dAll positive and negative counts are given for pairs \(\{\)probe, sample\(\}\)

ISSN: 1474-760X